So one can consider that another bug in the reader. It seems to work OK when it is working from a simple pathname (I have never seen GET/directory/filename.pdf#search='keyword'), but when it is reading from a script with parameter(s), it will just pass on the # probably thinking it is part of a parameter value.
That depends. It looks like at least some browsers send the # to the server when it is part of a parameter, not something that looks like a pathname.
On our website we have a directory with.pdf files. On the site there are two kinds of links to it:
1. of the form/directory/filename.pdf which return the content as application/pdf which normally results in an embedded reader window. intended to view the document.
2. of the form/directory?file=filename.pdf which is handled by an index.php in the directory that processes the ?file= arg (ofcourse checking that it does not try to refer a file outside of the directory) and sends the file with a Content-disposition: attachment; filename="filename.pdf" header, which normally results in a popup to save the file.
Now I have found that sometimes we get references to URL of the form/directory?file=filename.pdf#search='searchkeyword ' It turns out that this happens when the user has selected that second form (which is intended for downloading the pdf rather than viewing it), then still selected it to be "opened by the application" from the browser save-file popup, and then performed a search operation inside the reader. I found this a rather interesting form of information leakage, as you can see what search terms the users are using on your pdf... it must be intentional.
Maybe in this case that particular behaviour can be used by having URL of the form/directory?file=filename.pdf and the index.php (or.pl or whatever scripting language you like) sending just the mime type and the file, not the content disposition. Disadvantage of course is that the reader cannot load partial data as it normally does when it is reading from a file.
The display standard used by Videotex and Teletext was defined in 1974. The first services that used it appeared around 1977. That predates the development and release of the personal computers that you compare it with (which appeared end 1977 and in 1978).
Of course, larger scale rollout of videotex and teletext happened only by 1980, and by then it would have been done differently when development only had started at that date. A lot happened in those six years.
In those days we were all focussed on getting a reasonable performance out of the hardware we had to work with. A 6502 or Z-80 could scroll a 1-2 KB videoram in reasonable time (of course videotex does not use scrolling), but for the typical 12-16 KB of RAM required for a fully pixel-addressable display at TV resolution the processor was too slow to scroll it without painfully visible ripple. So most systems (including the ones you name) operated by default (or only) in a textmode with fixed width and a character generator. Full graphic mode was only used for special purposes, and even then there often was some effort made so you could use the graphic mode for graphics, and still have the character mode available for text overlays so the system did not have to resort to graphic character painting. "Painting" a fullsize graphic screen usually required a good fraction of a second in those days.
This is unfair critisism. In the days when Videotex was designed and rolled out, it was very rare to have a pixel-addressable graphic display, especially in an inexpensive device. All video displays of those days have a fixed arrangement of so-many-characters by so-many-lines, where the cells of the matrix are displayed using a "character generator", a lookup table that displays a certain character in a pixel grid within the character cell. This reduces the amount of memory, and also reduces the amount of work the software has to perform to fill the display. (it was much faster to put a single ASCII code in a memory cell to display a single character, than to copy the pixel image of that character into 8-16 locations in a graphic display)
The result of this design is that the characters are fixed-size. The "graphics" are actually a range of characters cleverly designed to form a complete set of all combinations of 2x3 blocks within the character cell. So the resolution is less than the pixel size, a result of the savings in memory.
The first personal computers available often had a similar (or the same) display technology. So it is not like they overlooked a capability that would have made it into a much better system.
The web was the same, before the "website designers" entered the scene. Many early webpages were plaintext, with only the occasional picture like a company logo or a nice-looking bullet.
Indeed, this was almost a requirement in the 14k4 modem days. It can still be done today, but apparently it does not sell in most markets.
Experience in the past has showed that we better wait until SP1 or maybe even SP2 before considering a migration. Maybe 10% of the workstations at work is "Vista ready". We most likely will not write off the other 90%.
I would say, give it at least a year and probably a servicepack, and only then start evaluating the costs and benefits.
What I mean is: when you see a CRT that is specified as 32", it will be a lot smaller than an LCD specified at 32". So when you want to compare screens of equal size, you will have to use measuring tape, not the spec sheet.
Also do not forget that the inches measuring up a CRT are smaller than those of an LCD screen. That 40" CRT probably is around the same size as a 37" LCD.
You really cannot compare CRT and LCD screen sizes.
I also live in Europe and when I wanted to buy a new TV two years ago I spent some time looking around. CRT still was quite present in the stores then. (right now it is mostly gone)
I looked at Philips, Sony, Panasonic, JVC etc etc but I was unable to find a 32/36" 16:9 set in the high-end segment that provided reasonable geometry and stability. Ironically, most stores have a music TV station on all the sets, and you can see the picture pump in and out (and the on-screen logo's jump around) in the rhytm of the brightness changes of the typical videoclip. Watching 4:3 made my eyes sore. You can select between true aspect ratio with two black sides that are more accurately described as curves, or distorted picture because of nonlinear stretching to 16:9.
I was quite disappointed because those sets would have to cost around 2000 euro and up, and even a low-end computer CRT monitor had so much better geometry and so much more possibility for alignment.
Finally, I bought an LCD TV. Perfect geometry and stability. 4:3 with perfect borders, no burn-in like on Plasma. In a very dark room the purple bleed-through of the LCD is visible, but in general the picture is very good, and much better than my old CRT. When the room is sun-lit during the summer, the bright and nonreflecting LCD image is especially good. (CRT and Plasma always have a lot of reflection in the screen glass)
Windows DST has several problems. One is the changing of the CMOS clock described in other followups.
Another is that there is only one start/end of changeover stored, so it does not remember in what years the DST has started at what time. Furthermore (related to this), when DST has started or ended, the filetime as seen in applications actually changes for files created long ago.
So, if during winter you create a file, e.g. at 14:00 in the afternoon, then you go back looking at it during summer, the time indicated no longer will be 14:00.
Linux does this much better. Every file gets a timestamp in UTC that gets converted to local time when it is displayed, and the conversion takes into account the DST rules and actual offset at the time indicated by the timestamp, so it does not arbitrarily change later.
I think Windows (in NTFS) stores UTC timestamps as well, they just have their UTC to local time functions goofed up. In FAT32 filesystems, stamps are in local time.
Chinnery says he's accepted the fact that he'll have to use the utility to fix his Windows 2000 systems. But, lacking an easily deployable patch, it means he must walk around to tweak each machine in his organization. This is a chore he doesn't feel he should face.
This is what you get for having systems that can be administered using a simple mouseclick by somone with only superficial knowledge of the matter! Of course it would be simple to automatically install a registry fix on all systems on his network, but he has become so accustomed to every tiny fix being installed in a hundreds of KB executable with automatic installer that he has never learned (or forgotten) how to script such simple things himself.
The daylight saving time mechanism in Windows is broken anyway. Posix DST handling is much better, especially (but not only) when the definition of start and end dates changes.
From the tuner on, the demodulated DVB bitstream is compatible with ATSC as well... Hell, from the tuner on, the demodulated DVB bitstream is compatible with a $20 DVD player...
No. DVB is not only MPEG2. It defines the stream multiplexing and timing, additional services like subtitling, teletext, auxiliary data, multiple sound channels, EPG, encryption, interactive services, etc etc etc.
All of these are defined based on the bitstream that comes out of the tuner. The only difference between the DVB versions is how this is modulated on a carrier, not how the bits are to be interpreted.
Look at it in the way you can have an ethernet card with coax, with UTP or with Fiber. The medium is different but the frames are the same, so the software does not need to know what medium is used. And you can have ethernet cards that support more than one medium.
I'm sure they'd love to travel back in time and build their satellite to standards that were nonexistant at the time, if they could.
Actually the satellites are not at all part of this. They just have transponders that relay the signals sent from earth. The same transponders that relayed analog TV were later used to relay digital TV. Their only problem to change standard could be an installed base of set-top boxes they don't want to replace.
The studies I have read on the Internet make a big fuss about a difference of a few dB in required transmitter power (the most I have seen is 6dB, or a factor of 4), and downplay the multipath problem. Here we have a DVB-T network that has all transmitters operating on the same channels. Small transmitters scattered over the area instead of one big tower with a megawatt ERP (as we had with analog TV). Try that with 8VSB.
For example, DVB covers Terrestrial, Satellite, and Cable, but you have to buy AN ENTIRELY DIFFERENT CARD FOR EACH ONE.
That is an implementation issue. The DVB systems use a different modulation optimized for each medium, and you need a separate tuner for each of them. However, the demodulated bitstream is the same and from the tuner on, the hardware is the same. Receivers are available that have their tuners seated in slots, and that can accept a user-selected mix of DVB-S, DVB-C and DVB-T tuners.
Well, this Vivanco box is quite unique in that it has RGB on all connectors, and it works bi-directionally. There are 5 SCART connectors and it autoswitches to the last unit that was powered on (it also has a remote). The selected input signal is output on all other connectors! I have a long SCART cable running to the bedroom TV, connected to one of the switchbox connectors. So in fact it is working as a 3-input 2-output switch for me.
The TV has side-composite and S-video, component and VGA/DVI/HDMI. It is one of those Ambilight LCD screens, very nice. In a very dark scene the contrast could be better, but other than that the picture is very good.
The DVB standard has a wider scope. The same digital signal is sent, using different modulation, over satelite, cable or terrestrial. Now even over mobile phone links. When the modulation would be a major concern, it would be possible to do DVB over the 8VSB modulation used by ATSC. But independent testing in countries wanting to adopt a digital TV standard often showed that OFDM (DVB-T) was superior to 8VSB (ATSC) in practice.
Probably because the second SCART had the S-Video connectors. It's either/or when it comes to SCART unfortunately.
This is incorrect. True that usually on a multi-SCART TV not all connectors can process all signals, but on my TV one of the SCARTs does both RGB and S-video. It is even autodetecting it. But I have seen other TVs that have to be configured for RGB or S-Video (Y/C) somewhere in a menu.
Last time I checked, it was not even available in Dutch. That may be one of the reasons it is not offered automatically here. Looking at the stats on the webserver at work, I see only 3% of MSIE 7 visitors. This means our visitors, which are mainly from the Netherlands, probably don't get this update pushed automatically. (MSIE 6 is at 78.7% and Windows XP at 68.1%)
Version 7 is clean -- which is welcome news in this first update since the upgrade was pushed to the world last month.
I know you Americans consider "the USA" the same as "the world", but I can assure you that IE7 was NOT pushed out in the Dutch version of Windows XP. It is not even available as an optional package in Windows update. And I think it is the same in many other countries.
Just buying a STB and hooking it up isn't enough for everyone - depending on coverage for your area you might need to spend money on your aerial.
On the switchover day, the existing analog transmission towers were fitted with digital transmitters sending the old analog programme package as a single digital multiplex. The coverage of these transmitters is roughly the same as the original analog signal (although the power is much lower). So, connecting the existing aerial used for analog reception should work.
In addition to this, there is a commercially operated DVB-T network being rolled out, which transmits from many lower locations in cities. The public channels on this network are now made free-to-air. The remainder of the package is scrambled and only available to subscribers. Depending on where you are, it may be possible to receive this network instead. However, it is vertically polarized instead of horizontal, and it is designed to be used with settop (indoor) antennas.
Our public TV system is only state-sponsored in the sense that it gets partly funded by tax money. There is no direct state influence on the programme content. The state only speficies the quality criteria for the total programme package.
Anyone can receive commercial programmes, but the broadcasters of those programmes have chosen to have the cost of distribution directly paid for by the viewers. So, you have to get a (cable, satellite or dvb-t) subscription to watch commercial TV. that is not a state decision, it could have been FTA if the costs were paid by the commercial channels (who would of course have to recoup that via even more advertising, like in the US).
i.e., virtually nobody used analogue television in the country anyway.
This probably was possible because the transmission of TV programmes has been a monopoly of the state for 35 years. Over that time we gradually got 3 national channels, transmitted in analog PAL.
When commercial TV was made possible (legally), satellite already was available as a distribution method and cable networks were available in all cities. So, no commercial TV provider was even considering building a TV transmitter network. They all transmitted exclusively on satellite and had their transmissions relayed on cable. (we have about 10 commercial channels)
So, everyone who wanted to watch more than the state TV (equiv to BBC) already had to have either cable or satellite. In the UK, the situation is quite different as commercial channels like ITV started earlier and could be received in terrestrial analog.
Spanish channels on Hispasat (30W) transmit hard-core porn on free-to-air channels between 00:30 and 05:00.
It is not clear to me under what regulation they do that, as I believe it is not completely legal to do so in the EU. (of course there are too many soft channels to count)
Not only that. In France, there was an 819 line system (today, we would probably call that HDTV) that was replaced by 625 line SECAM when color was introduced.
(the remainder of the continent mostly used 625 line systems and went with 625 line PAL, which was compatible with the B/W system)
Slashdot Mirror
So one can consider that another bug in the reader. It seems to work OK when it is working from a simple pathname (I have never seen GET /directory/filename.pdf#search='keyword'), but when it is reading from a script with parameter(s), it will just pass on the # probably thinking it is part of a parameter value.
That depends. It looks like at least some browsers send the # to the server when it is part of a parameter, not something that looks like a pathname.
.pdf files. On the site there are two kinds of links to it:
/directory/filename.pdf which return the content as application/pdf which normally results in an embedded reader window. intended to view the document.
/directory?file=filename.pdf which is handled by an index.php in the directory that processes the ?file= arg (ofcourse checking that it does not try to refer a file outside of the directory) and sends the file with a Content-disposition: attachment; filename="filename.pdf" header, which normally results in a popup to save the file.
/directory?file=filename.pdf#search='searchkeyword '
/directory?file=filename.pdf and the index.php (or .pl or whatever scripting language you like) sending just the mime type and the file, not the content disposition.
On our website we have a directory with
1. of the form
2. of the form
Now I have found that sometimes we get references to URL of the form
It turns out that this happens when the user has selected that second form (which is intended for downloading the pdf rather than viewing it), then still selected it to be "opened by the application" from the browser save-file popup, and then performed a search operation inside the reader.
I found this a rather interesting form of information leakage, as you can see what search terms the users are using on your pdf... it must be intentional.
Maybe in this case that particular behaviour can be used by having URL of the form
Disadvantage of course is that the reader cannot load partial data as it normally does when it is reading from a file.
The display standard used by Videotex and Teletext was defined in 1974. The first services that used it appeared around 1977.
That predates the development and release of the personal computers that you compare it with (which appeared end 1977 and in 1978).
Of course, larger scale rollout of videotex and teletext happened only by 1980, and by then it would have been done differently when development only had started at that date. A lot happened in those six years.
In those days we were all focussed on getting a reasonable performance out of the hardware we had to work with. A 6502 or Z-80 could scroll a 1-2 KB videoram in reasonable time (of course videotex does not use scrolling), but for the typical 12-16 KB of RAM required for a fully pixel-addressable display at TV resolution the processor was too slow to scroll it without painfully visible ripple.
So most systems (including the ones you name) operated by default (or only) in a textmode with fixed width and a character generator. Full graphic mode was only used for special purposes, and even then there often was some effort made so you could use the graphic mode for graphics, and still have the character mode available for text overlays so the system did not have to resort to graphic character painting.
"Painting" a fullsize graphic screen usually required a good fraction of a second in those days.
This is unfair critisism. In the days when Videotex was designed and rolled out, it was very rare to have a pixel-addressable graphic display, especially in an inexpensive device.
All video displays of those days have a fixed arrangement of so-many-characters by so-many-lines, where the cells of the matrix are displayed using a "character generator", a lookup table that displays a certain character in a pixel grid within the character cell. This reduces the amount of memory, and also reduces the amount of work the software has to perform to fill the display.
(it was much faster to put a single ASCII code in a memory cell to display a single character, than to copy the pixel image of that character into 8-16 locations in a graphic display)
The result of this design is that the characters are fixed-size. The "graphics" are actually a range of characters cleverly designed to form a complete set of all combinations of 2x3 blocks within the character cell. So the resolution is less than the pixel size, a result of the savings in memory.
The first personal computers available often had a similar (or the same) display technology. So it is not like they overlooked a capability that would have made it into a much better system.
The web was the same, before the "website designers" entered the scene.
Many early webpages were plaintext, with only the occasional picture like a company logo or a nice-looking bullet.
Indeed, this was almost a requirement in the 14k4 modem days. It can still be done today, but apparently it does not sell in most markets.
Experience in the past has showed that we better wait until SP1 or maybe even SP2 before considering a migration.
Maybe 10% of the workstations at work is "Vista ready". We most likely will not write off the other 90%.
I would say, give it at least a year and probably a servicepack, and only then start evaluating the costs and benefits.
What I mean is: when you see a CRT that is specified as 32", it will be a lot smaller than an LCD specified at 32".
So when you want to compare screens of equal size, you will have to use measuring tape, not the spec sheet.
Also do not forget that the inches measuring up a CRT are smaller than those of an LCD screen.
That 40" CRT probably is around the same size as a 37" LCD.
You really cannot compare CRT and LCD screen sizes.
I also live in Europe and when I wanted to buy a new TV two years ago I spent some time looking around. CRT still was quite present in the stores then.
(right now it is mostly gone)
I looked at Philips, Sony, Panasonic, JVC etc etc but I was unable to find a 32/36" 16:9 set in the high-end segment that provided reasonable geometry and stability.
Ironically, most stores have a music TV station on all the sets, and you can see the picture pump in and out (and the on-screen logo's jump around) in the rhytm of the brightness changes of the typical videoclip.
Watching 4:3 made my eyes sore. You can select between true aspect ratio with two black sides that are more accurately described as curves, or distorted picture because of nonlinear stretching to 16:9.
I was quite disappointed because those sets would have to cost around 2000 euro and up, and even a low-end computer CRT monitor had so much better geometry and so much more possibility for alignment.
Finally, I bought an LCD TV. Perfect geometry and stability. 4:3 with perfect borders, no burn-in like on Plasma.
In a very dark room the purple bleed-through of the LCD is visible, but in general the picture is very good, and much better than my old CRT. When the room is sun-lit during the summer, the bright and nonreflecting LCD image is especially good.
(CRT and Plasma always have a lot of reflection in the screen glass)
Windows DST has several problems.
One is the changing of the CMOS clock described in other followups.
Another is that there is only one start/end of changeover stored, so it does not remember in what years the DST has started at what time.
Furthermore (related to this), when DST has started or ended, the filetime as seen in applications actually changes for files created long ago.
So, if during winter you create a file, e.g. at 14:00 in the afternoon, then you go back looking at it during summer, the time indicated no longer will be 14:00.
Linux does this much better. Every file gets a timestamp in UTC that gets converted to local time when it is displayed, and the conversion takes into account the DST rules and actual offset at the time indicated by the timestamp, so it does not arbitrarily change later.
I think Windows (in NTFS) stores UTC timestamps as well, they just have their UTC to local time functions goofed up.
In FAT32 filesystems, stamps are in local time.
Chinnery says he's accepted the fact that he'll have to use the utility to fix his Windows 2000 systems. But, lacking an easily deployable patch, it means he must walk around to tweak each machine in his organization. This is a chore he doesn't feel he should face.
This is what you get for having systems that can be administered using a simple mouseclick by somone with only superficial knowledge of the matter!
Of course it would be simple to automatically install a registry fix on all systems on his network, but he has become so accustomed to every tiny fix being installed in a hundreds of KB executable with automatic installer that he has never learned (or forgotten) how to script such simple things himself.
The daylight saving time mechanism in Windows is broken anyway. Posix DST handling is much better, especially (but not only) when the definition of start and end dates changes.
From the tuner on, the demodulated DVB bitstream is compatible with ATSC as well... Hell, from the tuner on, the demodulated DVB bitstream is compatible with a $20 DVD player...
No. DVB is not only MPEG2. It defines the stream multiplexing and timing, additional services like subtitling, teletext, auxiliary data, multiple sound channels, EPG, encryption, interactive services, etc etc etc.
All of these are defined based on the bitstream that comes out of the tuner. The only difference between the DVB versions is how this is modulated on a carrier, not how the bits are to be interpreted.
Look at it in the way you can have an ethernet card with coax, with UTP or with Fiber. The medium is different but the frames are the same, so the software does not need to know what medium is used.
And you can have ethernet cards that support more than one medium.
I'm sure they'd love to travel back in time and build their satellite to standards that were nonexistant at the time, if they could.
Actually the satellites are not at all part of this. They just have transponders that relay the signals sent from earth. The same transponders that relayed analog TV were later used to relay digital TV. Their only problem to change standard could be an installed base of set-top boxes they don't want to replace.
The studies I have read on the Internet make a big fuss about a difference of a few dB in required transmitter power (the most I have seen is 6dB, or a factor of 4), and downplay the multipath problem. Here we have a DVB-T network that has all transmitters operating on the same channels. Small transmitters scattered over the area instead of one big tower with a megawatt ERP (as we had with analog TV). Try that with 8VSB.
For example, DVB covers Terrestrial, Satellite, and Cable, but you have to buy AN ENTIRELY DIFFERENT CARD FOR EACH ONE.
s _dm7025.php
g .pdf
That is an implementation issue. The DVB systems use a different modulation optimized for each medium, and you need a separate tuner for each of them.
However, the demodulated bitstream is the same and from the tuner on, the hardware is the same.
Receivers are available that have their tuners seated in slots, and that can accept a user-selected mix of DVB-S, DVB-C and DVB-T tuners.
E.g. this one: http://www.dream-multimedia-tv.de/english/product
ATSC also has (long had) standards for broadcast, cable, and satellite transmission.
Yet on American DTH satellites, DVB is used by one distributor and DSS by another.
Sources?
http://www.ebu.ch/en/technical/trev/trev_295-huan
Well, this Vivanco box is quite unique in that it has RGB on all connectors, and it works bi-directionally.
:-(
There are 5 SCART connectors and it autoswitches to the last unit that was powered on (it also has a remote).
The selected input signal is output on all other connectors! I have a long SCART cable running to the bedroom TV, connected to one of the switchbox connectors.
So in fact it is working as a 3-input 2-output switch for me.
The TV has side-composite and S-video, component and VGA/DVI/HDMI. It is one of those Ambilight LCD screens, very nice. In a very dark scene the contrast could be better, but other than that the picture is very good.
My table is covered with remotes.
My TV is about 2 years old. It is a Philips. I think the autodetect was somehow motivated by their "Sense & Simplicity" marketing.
It has 3 SCART inputs. One with RGB, one with RGB and Y/C and one with only composite.
This still isn't enough for me, so I have a separate SCART-RGB switchbox (Vivanco AV Control 5) that I use to switch between 3 SCART-RGB sources.
The DVB standard has a wider scope. The same digital signal is sent, using different modulation, over satelite, cable or terrestrial. Now even over mobile phone links.
When the modulation would be a major concern, it would be possible to do DVB over the 8VSB modulation used by ATSC.
But independent testing in countries wanting to adopt a digital TV standard often showed that OFDM (DVB-T) was superior to 8VSB (ATSC) in practice.
Probably because the second SCART had the S-Video connectors. It's either/or when it comes to SCART unfortunately.
This is incorrect. True that usually on a multi-SCART TV not all connectors can process all signals, but on my TV one of the SCARTs does both RGB and S-video.
It is even autodetecting it. But I have seen other TVs that have to be configured for RGB or S-Video (Y/C) somewhere in a menu.
Being interested in the US IT industry or US IT news is not the same as equating "the US" to "the world"...
Last time I checked, it was not even available in Dutch. That may be one of the reasons it is not offered automatically here.
Looking at the stats on the webserver at work, I see only 3% of MSIE 7 visitors. This means our visitors, which are mainly from the Netherlands, probably don't get this update pushed automatically.
(MSIE 6 is at 78.7% and Windows XP at 68.1%)
Version 7 is clean -- which is welcome news in this first update since the upgrade was pushed to the world last month.
I know you Americans consider "the USA" the same as "the world", but I can assure you that IE7 was NOT pushed out in the Dutch version of Windows XP. It is not even available as an optional package in Windows update.
And I think it is the same in many other countries.
Just buying a STB and hooking it up isn't enough for everyone - depending on coverage for your area you might need to spend money on your aerial.
On the switchover day, the existing analog transmission towers were fitted with digital transmitters sending the old analog programme package as a single digital multiplex.
The coverage of these transmitters is roughly the same as the original analog signal (although the power is much lower).
So, connecting the existing aerial used for analog reception should work.
In addition to this, there is a commercially operated DVB-T network being rolled out, which transmits from many lower locations in cities. The public channels on this network are now made free-to-air. The remainder of the package is scrambled and only available to subscribers.
Depending on where you are, it may be possible to receive this network instead. However, it is vertically polarized instead of horizontal, and it is designed to be used with settop (indoor) antennas.
Our public TV system is only state-sponsored in the sense that it gets partly funded by tax money. There is no direct state influence on the programme content.
The state only speficies the quality criteria for the total programme package.
Anyone can receive commercial programmes, but the broadcasters of those programmes have chosen to have the cost of distribution directly paid for by the viewers.
So, you have to get a (cable, satellite or dvb-t) subscription to watch commercial TV. that is not a state decision, it could have been FTA if the costs were paid by the commercial channels (who would of course have to recoup that via even more advertising, like in the US).
i.e., virtually nobody used analogue television in the country anyway.
This probably was possible because the transmission of TV programmes has been a monopoly of the state for 35 years.
Over that time we gradually got 3 national channels, transmitted in analog PAL.
When commercial TV was made possible (legally), satellite already was available as a distribution method and cable networks were available in all cities.
So, no commercial TV provider was even considering building a TV transmitter network. They all transmitted exclusively on satellite and had their transmissions relayed on cable. (we have about 10 commercial channels)
So, everyone who wanted to watch more than the state TV (equiv to BBC) already had to have either cable or satellite.
In the UK, the situation is quite different as commercial channels like ITV started earlier and could be received in terrestrial analog.
Spanish channels on Hispasat (30W) transmit hard-core porn on free-to-air channels between 00:30 and 05:00.
It is not clear to me under what regulation they do that, as I believe it is not completely legal to do so in the EU.
(of course there are too many soft channels to count)
Not only that. In France, there was an 819 line system (today, we would probably call that HDTV) that was replaced by 625 line SECAM when color was introduced.
(the remainder of the continent mostly used 625 line systems and went with 625 line PAL, which was compatible with the B/W system)